Review of “Spectral optical properties of soot: laboratory investigation of propane flame particles and their link to composition”  

By

Johannes Heuser, Claudia Di Biagio, Jerome Yon, Mathieu Cazaunau, Antonin Bergé, Edouard Pangui, Marco Zanatta, Laura Renzi, Angela Marinoni, Satoshi Inomata, Chenjie Yu, Vera Bernardoni, Servanne Chevaillier, Daniel Ferry, Paolo Laj, Michel Maillé, Dario Massabò, Federico Mazzei, Gael Noyalet, Hiroshi Tanimoto, Brice Temime-Roussel, Roberta Vecchi, Virginia Vernocchi, Paola Formenti, Bénédicte Picquet-Varrault, and Jean-François Doussin 

General comments 

The optical properties of atmospheric soot aerosols are crucial for understanding the role they play in the climate system. Soot has been deemed a vital part of the climate system because of its ability to affect the vertical temperature structure of the atmosphere, affect cloud lifetime, and how soot aerosols can reduce the albedo of e.g. snow once they are removed from the atmosphere through deposition. 

The manuscript describes results from well controlled chamber experiments on soot aerosol particles generated with a miniCAST burner during an extensive campaign using the CESAM chamber. The chamber allows for long experiments due to its 4.2 m2 volume. The study describes data obtained using five different setpoints of the miniCAST burner to produce soot aerosol particles (cast soot, abbreviated as CS in the manuscript) of different chemical composition and physical size. The soot aerosols range from mostly elemental carbon (CS1) to mostly organic carbon (OC) in the CS5 experiment.  

The data analysis is for the most part sound and appears to be well executed. The figures are clear and informative. 

My main concern is whether these experiments are relevant for ambient aerosols since all the experiments (CS1-CS5) show soot that is still in it’s fractal state. A fractal state is expected since the experiments were conducted under dry and dark conditions on fresh soot. In the atmosphere, soot cores will collapse if they are hydrophilic, and that will change the aerosols optical properties (see. e.g. https://doi.org/10.1029/2005JD006389). Collapsed fractal aggregates will have very different optical properties (see e.g. https://doi.org/10.1029/1998JD100069) and thus (MAC, MSC, AAE, and SSA). The question remains, how well these CS1-CS5 cases represent ambient aerosol particles and what conclusions can be drawn from these experiments that are relevant for our understanding of ambient aerosols or to be utilized in models.  

The shortcoming of the manuscript becomes evident in the introduction when the authors states that the study “...aim to measure...” and “tries to identify generalized tendencies ... on soot composition.” First, vague wordings do not invite further reading. Second, and more importantly, in the introduction it is not stated what is new to science in this study, which would raise the reader’s interest. I suggest the authors make a greater effort to highlight what added information comes out from these experiments and how it helps to increase our understanding about atmospheric aerosols. The results based on Figure 7 should be put as a goal of this study and be mentioned in the introduction. 

I suggest changing the manuscript's focus on what you know about the aerosols you have generated, thus leaving the speculation related to CS5 to a separate section of the manuscript. This would make for easier and more interesting reading. As it reads now, much of the results are diluted by speculation about the CS5 aerosols. Speculation about CS5 draws too much attention away from the well-defined experiments of CS1-CS4. There is also a mention of an ACSM being present during the experiments, but the data is never properly used. If you are able to extend the analysis to provide more information about the CS5 aerosol particles, and reduce the amount of speculation, please do so. 

I like the work you have done relating to Figure 7a and 7b where you have put your research in context with previous research on the topic and included other types of soot aerosol than the miniCAST soot. When reading the manuscript, it becomes clear that the authors have a vast knowledge about previous literature on the subject. In Fig. 7, the authors have made good use of this knowledge. I would like to encourage the authors to do so more often. As it reads now, the interpretation of the results in relation to other studies is often left to the reader, with a list of articles to read and interpret for themselves. I’ve listed a few examples below: 

P17 L491-493: “The diameter sits at the lower end of the primary spheres observed in combustion aerosol (5 refereces)” It would be much more valuable to have the authors analyze the result than give a list of articles.  

P17 L494-495: “This value is located towards the upper end of observed values (3 references)” What is the range in those articles? 

P18 502-504: “The Dfm is our study is in line with previous studies on propane miniCAST and ethylene soot (4 references), while it is on the low end for values observed for diesel and aircraft soot (5 references)” Can the authors put some numbers on these differences and similarities so that the reader won’t have to read all the listed publications for themselves. That would give much added value to your work than to just cite others.  

I would further suggest the authors try to explain what these experiments mean for our understanding of ambient aerosols. The authors have already stated that in the introduction of the manuscript in terms of MAC, MSC, MEC, and SSA of black carbon (BC) containing aerosol particles; i.e. how can your findings be utilized in models or measurements of atmospheric aerosols. I have made a few suggestions further on.  

The aerosols, as the authors correctly state, are representative of freshly emitted aerosols. The conditions in the chamber where dark and dry, which is not a condition that one is likely to come across in the atmosphere, at least considering the longer chamber residence times in the chamber which were up to 27 h. It is not clear to me why these conditions were chosen although I realize that these conditions make for an easier data analysis and well constrained conditions. I would like to see the motivation for this and discussion about how the aerosols you have generated can be useful for understanding soot in ambient conditions.  

My suggestion is that you include discussion on what happens when these fractal like soot aggregates enter the atmosphere and that your values and findings apply to legacy soot up to a certain point, before they collapse into more compacted shapes due to atmospheric processing, after which the optical properties you report no longer holds true. My suggestion is that you include discussion on at which stage the soot cores will collapse; see https://doi.org/10.1016/j.carbon.2024.119197 and https://doi.org/10.1038/ngeo2901 and references therein. Then you could state that your experiments can be used in models up to a certain point (i.e. before they collapse). Based on your synthesis in the discussions section you could also recommend setpoints for the miniCAST to replicate soot types that are found in the ambient. That information is in the manuscript (and in the references) already but is not explicitly stated.  

The strength of this work is in the discussions section of the manuscript where the experiments are put into context with previous research and other types of generated aerosols. Before reaching the discussions section, one has to read quite a lot of details about the experiments. I would not mind if the authors could review if everything in sections 2 and 3 is needed in the revised manuscript, keeping in mind that something could be put into the supplements, such as e.g. Fig. 2 and Fig. 6.  

Specific comments 

P3 L75-78: I don’t understand the reasoning here. Light scattering by aerosol particles is mainly due to the size of the particles and their refractive index, whereas light absorption is due to the amount of BC in the aerosol. In the atmosphere, BC is often coated or at least attached to a non-absorbing material. If extinction is dominated by absorption then it means that the particle is mainly BC or so small in size that it does not scatter light. Why not talk about single scattering albedo (SSA) here instead?  

P4 L103: There is no mention of the size dependence of MAC, MSC, MEC and SSA in the introduction. The primary spherules of the fractal aggregates (shown in Figure 3) are surely in the Rayleigh regime and are therefore volume absorbers (https://doi.org/10.3155/1047-3289.59.9.1028). Light scattering is proportional to the optical size of the particles; see the size parameter x in Fig. 1 of https://doi.org/10.3155/1047-3289.59.9.1028. I suggest to also include the SAE when reporting MSC in the article as the SAE is a measure of the optical size of the particles. This information is included in Table 2, but not in the text in the results/discussion section to the degree that it would become clear that the MSC is indeed largely due to the size of the particles.  

P4 L111: What single soot particles? The sentence is too long and raises more questions than it answers. A lot of references to BC-like soot particles and an ongoing discussion but I would urge the authors to provide your thoughts and summary on this issue as few readers know all these publications by heart and immediately know what you mean.  

P7 L216-220: Where is this ACSM data used? This data could be useful when trying to understand the CS5 experiment in greater detail. On P15

L442-445 there is a mention of this data, but the reliability of the results is put into question.  

P11L332-335: Please revisit this sentience as it is a bit unclear what is meant here.  

P14 L407: It is not clear to me why the distributions of Fig. 4 should be gaussian. If the aerosols in the chamber is not monomodal, or if the particle size distribution changes due to e.g. coagulation, I don’t see reason for why a gaussian fit is the best matrix for the cross section statistics. For some experiments it works suprisingly well, but for some I’m sure that e.g. a mean value would make more sense than to report the peak of the fit. Please, consider to at least report both. The figure itself is informative and supports the text and overall message that the experiments were successful and well constrained.  

P15 L435-438: Please try and make this sentence more clear. 

P15 L439-447: As I suggested before, I would like to see the CS5 experiment separated from the CS1-4 experiments since it dilutes the discussion of CS1-4 too much.  

P21 L578: What is meant with “over particle lifetime”? That OC particles in the Aitken mode coagulate with the larger soot particles in the chamber? 

P22 L606-609: Can the authors recommend setpoints for the miniCAST to reproduce some of these aerosol types that are found in the ambient air? This would be valuable for the scientific community. 

P22 L610: Remove “Relating the...” from the heading. 

P23 L615-617: Please split this one sentence into more sentences to make the message clear. Now the message is too convoluted. Isn’t the spectral dependence of absorption (please use AAE) depending on the mixture of fuel, oxidation air and N2 and not the “absolute” value of these? 

P24 Figure 7: I think panels (a) and (b) in Figure 7 are very informative and a great addition to literature and potentially useful for modelers. Having said that, I don’t see the point of panels (c) and (d). The SAE and MSC are mainly a  function of particle size and has little to do with the OC/EC ratio. Surely the OC/EC ratio will change the refractive index (and density) of the partilces and thus also impact the SAE and MSC, but SAE will still be dictated by the size of the particles (and MSC by the density and size of the particles).  

P25 L697-P26 L698: “the SAE suggest a potential similar relation between composition and spectral dependence of scattering as observed for the absorption” This is because the particle size changed between the experiments CS1-CS5 (CMD in Table 1). The spectral dependence of scattering (please use SAE) depends primarily on particle size and not the OC/EC ratio.  

P26 L710: MSC is primarily a function of particle size and particle density, and not EC/TC content. 

Technical corrections (P = page, L = line) 

P2 L41: “...during the combustion ...” remove “the” 

P2 L45: “laser incandescence” add induced 

P4 L131: remove “realism” 

P5 L134: remove “tries” 

P8 L258: change “cast soot” to CS. Remove “around“ 

P11 L342: Please change to “190 to 640 nm wavelength range” 

P21 L 564: BrCACN is not defined yet. 

P22 L581: wavelength dependency should be AAE? 

P22 L583: Change “could” to can. 

P22 L588: “Seems to” is rather vague. 

P22 L599-600 Please rephrase “such a less” 

P22 L601: Please remove “As a matter of fact”. 

P23 L 628: Please remove “in particular”.  

P23 L 631: change “transformed” to calculated.  

Figure 6. Label the panels (a) and (b). 

Figure 7. Schnaiter et al. 2003 and Ess et al. 2019 are too similar in color, and so is Kumar et al. 2018. 

Supplement: 

P2 L14: Why is this an alternative? I suggest just saying that it is the same as Fig. 4, but using the 630 nm wavelength.  

P3 Figure S3: Same as for Figure 7 panes (c) and (d) do not make sense to me so I would remove them. You could try and plot SAE vs MSC for your experiment. Schnaiter et al. 2003 and Ess et al. 2019 are very similar in color, and so is Kumar et al. 2018. 

P4 Figure S5: The blue wavelengths seems to match well but red (630 nm) shows higher extinction although both instruments should show the same value. Is this a calibration issue or can it be because they were not connected “simultaneously”. I’m not sure what it means that they were not connected simultaneously. The clock of the logging software should have been the same. Or is it a delay in the instruments or in the sampling lines.  

P5 Table 1: It would be very interesting to see the results of the chemically aged experiments in this manuscript. Especially how the optical properties and the TEM images of the CS aerosols from those experiments. If the soot cores collapsed during those experiments, that would provide much sought after information about how the CS aerosols you have generated would act if subject to more ambient like conditions, which the manuscript now lacks. The scientific impact of this paper would substantially improve by including the aged CS aerosol experiments. 

P6 Table 2: You might want to include the size range of the SMPS in the manuscript for easier reading. 

P8 change “Pattenuation” to attenuation. 

References 

https://doi.org/10.3155/1047-3289.59.9.1028 

https://doi.org/10.1029/2005JD006389 

https://doi.org/10.1029/1998JD100069 

https://doi.org/10.1016/j.carbon.2024.119197 

https://doi.org/10.1038/ngeo2901